No.
Vibe coding was not coined by OpenAi.
I will give credit to Karpathy, but I don't think he came up with it. He said it out loud though, and established it, or in my opinion made it mainstream in 2025. He took ownership, and I have seen this happen timeously in business. The term "vibe coding" did get to be promotionally used by OpenAI. They, however, didn't get to coin "AI Pair Programming", that was GitHub (using Codex). OpenAI did make their DevDay 2025 about vibe coding. Public sentiment has since made them reconsider how to navigate the negativity. Vibe coding implies that anyone could potentially be a developer. It hasn't stopped them, and others, using it in their marketing language. I will stick to what I said, the term "vibe coding" was vibing around OpenAI. I mean, how much time was there between Karpathy leaving OpenAI and posting that on X?
Vibe coding is literally someone just letting the AI build the whole thing and who cares what the code looks like. As long as it works. Whatever.
This largely depends on its vehicle. Many don't see it like that. I will tell you now that there are people who are vibe coding, who will tell you that they are being AI assisted.
Theo has it well summarised in this video:
Yes, I am aware, there are people who don't like Theo.
Assisted AI coding is getting agents to help with very specific areas of development and not make all the decisions.
Each with their own thought, and these philosophies will take more shape as time progresses. I am a traditionalist, so I understand the dissidence, but I am aware where to accept improvements. More work, less time.
Not AI in coding, but AI still get things wrong. I am not a programmer, I am a layman.
Vibe coding is fine for small simple apps and non critical things. But if you want robust, production grade software - you don’t vibe code it in a day. You get developers to build it - who may or may not use AI assistance.
Yes, I agree.
Just see what Karpathy said recently:
I've never felt this much behind as a programmer. The profession is being dramatically refactored as the bits contributed by the programmer are increasingly sparse and between. I have a sense that I could be 10X more powerful if I just properly string together what has become available over the last ~year and a failure to claim the boost feels decidedly like skill issue. There's a new programmable layer of abstraction to master (in addition to the usual layers below) involving agents, subagents, their prompts, contexts, memory, modes, permissions, tools, plugins, skills, hooks, MCP, LSP, slash commands, workflows, IDE integrations, and a need to build an all-encompassing mental model for strengths and pitfalls of fundamentally stochastic, fallible, unintelligible and changing entities suddenly intermingled with what used to be good old fashioned engineering. Clearly some powerful alien tool was handed around except it comes with no manual and everyone has to figure out how to hold it and operate it, while the resulting magnitude 9 earthquake is rocking the profession. Roll up your sleeves to not fall behind.
Then in response to an LLM (Opus 4.5) being good:
It’s very good. People who aren’t keeping up even over the last 30 days already have a deprecated world view on this topic.
This one:
Q: Do you feel these skills are substantial and transferrable to the future? Does someone who started 1-2 years ago have an advantage over someone who started a month ago? Will lessons today be useful 1-2 years ago? I feel like the "missing out" factor is still quite low.
A: Very good questions imo experienced devs have a real advantage but only if they rapidly progress through their grief cycle and adapt, now and onwards. Categorically rejecting or ignoring the new layer would be a mistake.
What is this now, vibe coding or AI-assisted programming? There are solutions being built to hold and operate these layers.
There are many more examples that I can pick on.
This isn't limited to coding/programming. I see this in ITIL, all reporting, documentation, etc. There are people who have the experience, but who are not reviewing and assessing the outputs. There is this rush to cut time. More and more tools that I am using are becoming guided with agentic implementations, and more times than not it behaviourly adds time.
Anyhow, this is as much as I will add to the topic. Someone wants to build a WordPress site.